If you start writing “how did Target” on Google, it fills in the rest for you: “..know a girl was pregnant before her father did?” It is clearly one of the landmark cases for data driven target marketing. Firstly, it shows that using data to predict behaviour is not sci-fi. Secondly, it teaches a lesson for future data-based marketing: be transparent and make the data useful for the customer.
Andrew Pole, a statistician, started working for retail company Target in 2002. His job was to find patterns of behaviour in shopping that would allow better targeted ads. Because customer habits are hard to break, there was a need to find a right time for changing behaviours. When people have their first child, they start buying new things. The difficulty for retail is that after the birth parents are overwhelmed by baby-related ads. Target wanted to figure out who was pregnant, not recently given birth.
Target uses specific code for most of its random customers. The guest ID number is given whenever possible and linked to credit cards, coupons, mails and refunds. According to a reporter of New York Times, guest ID is enriched with demographic information:
“…like your age, whether you are married and have kids, which part of town you live in, how long it takes you to drive to the store, your estimated salary, whether you’ve moved recently, what credit cards you carry in your wallet and what Websites you visit. Target can buy data about your ethnicity, job history, the magazines you read, if you’ve ever declared bankruptcy or got divorced, the year you bought (or lost) your house, where you went to college, what kinds of topics you talk about online, whether you prefer certain brands of coffee, paper towels, cereal or applesauce, your political leanings, reading habits, charitable giving and the number of cars you own.”
The father in the story had found out his daughter was receiving ads for baby items and went to question the manager in-store. The manager had no clue why this was happening and apologised.
The father later found out that there had been some “activity” in his house he was unaware of.
Target “knew” before because it had gathered data from the baby shower gift registry. The key connection was odourless lotion (because pregnancy makes your sense of smell stronger). You would also use more supplements like zinc and magnesium. Pole’s computers were able to identify 25 products and use them to make a pregnancy prediction score for any customer.
The company responded with Target declined to specify what demographics they use. They refused to speak with reporter Charles Duhigg and banned him from visiting the headquarters. Apparently Andrew Pole was also told not to speak about his work to reporters. It seemed like it didn’t occur to him that his work is a secret.
Target noticed that sending pregnancy-exclusive ads creeps people out. So they blend them with random-looking products, putting pacifiers next to lawn-moners. It likely has contributed to company’s growing revenues. “Soon after the new ad campaign began, Target’s Mom and Baby sales exploded”, Duhigg writes.
The future of target marketing might and should look different. The more people learn about how their data is used, the more they expect companies to explain how it is useful. Recommendations should be transparent and based on interaction rather than intrusive and creepy. Fine-grained and personal data outshines statistical predictions.
It’s worth noting we have a fear bias. We are more afraid of our data ending up in the wrong place than it not ending up in the right place. In 21st century, we probably cannot protect all individuals from all risks without a sacrifice in data-driven value creation. We can however start thinking the use of personal data as a mutual benefit.